- openSUSE ARM

[opensuse-arm] xf86-video-omap in Factory
by Guillaume Gardet 10 Jul '12

10 Jul '12

Hi, For information, xf86-video-omap is now part of factory: https://build.opensuse.org/package/show?package=xf86-video-omap&project=ope… This driver is needed to use omapdrm with Xorg. Alex, you can now remove xorg-x11-drv-omap from Factory:ARM and use instead xf86-video-omap. Maybe we need to update some images? Guillaume -- To unsubscribe, e-mail: opensuse-arm+unsubscribe(a)opensuse.org To contact the owner, e-mail: opensuse-arm+owner(a)opensuse.org

1 0

[opensuse-arm] omapdrm
by Alexander Graf 04 Jul '12

04 Jul '12

Howdy, During the past few days I've been trying to get omapdrm to talk to me. Thanks to Rob's help I finally got things working on 3.4 [1]. Something's still wrong with the pixel clock, so I can only use low (1280x1024) resolutions. And for 2D acceleration we'd need a binary blob. But overall, we can completely replace omapfb with omapdrm and have future proof gfx on OMAP! I pushed the relevant changes to the openSUSE-12.2 branch of the SUSE kernel repo. So the next time we get a kernel submission for Factory / 12.2, it should contain a working omapdrm driver. Just modprobe it and you're ready to go with a console. The great thing about it though is that we finally have a working Xorg driver that doesn't randomly black screen you and which actually support xrandr, so you can switch resolutions and connections on the fly. To try it out, just use the xorg-x11-drv-omap package from openSUSE:Factory:ARM. The name is not final yet, so expect it to change. Keep up the great work guys! We're getting closer to a usable 12.2 every single day :). Alex [1] http://csgraf.de/tmp/beagle-xm-3.4.jpg -- To unsubscribe, e-mail: opensuse-arm+unsubscribe(a)opensuse.org To contact the owner, e-mail: opensuse-arm+owner(a)opensuse.org

2 3

[opensuse-arm] xf86-video-omap to X11:XOrg and Factory?
by Guillaume Gardet 03 Jul '12

03 Jul '12

Hi, now that compile error is fixed, my xf86-video-omap package builds fine for armv7. Could I push it to X11:XOrg (and then Factory)? Any comments, or questions? Project: https://build.opensuse.org/package/show?package=xf86-video-omap&project=hom… Cheers, Guillaume -- To unsubscribe, e-mail: opensuse-arm+unsubscribe(a)opensuse.org To contact the owner, e-mail: opensuse-arm+owner(a)opensuse.org

2 1

[opensuse-arm] omap framebuffer DVI output: update for 3.5-rc*
by Guillaume Gardet 03 Jul '12

03 Jul '12

Hi, Some people may have notice that dvi output was no more working on beagleboard with recent git kernel. I succeeded to have DVI working with framebuffer driver (omapfb) on my beagleboard xM. Since 3.5-rc* you need to modprobe panel-tfp410 before omapfb: modprobe panel-tfp410 modprobe omapfb mode=dvi:1440x900MR-24@60 I updated the wiki accordingly: http://en.opensuse.org/openSUSE:Beagleboard_(xM) Guillaume -- To unsubscribe, e-mail: opensuse-arm+unsubscribe(a)opensuse.org To contact the owner, e-mail: opensuse-arm+owner(a)opensuse.org

1 0

[opensuse-arm] parted and x86 bootinfo in the MBR
by Alexander Graf 03 Jul '12

03 Jul '12

Hi Petr, When running parted to create or modify an MBR label, it tries to be clever and throws a dummy x86 boot sector onto the MBR, so the disk is bootable on PCs. While that's nice and helpful for a bunch of use cases, it doesn't exactly help when running on ARM. Some ARM devices, such as the OMAP3 series, have a buggy boot rom that simply goes into nirvana if it finds non-0 in the first 4 bytes of the SD card. Please see libparted/labels/dos.c:1271 for the place that copies a generic x86 asm MBR into the MBR when its first byte is 0. We need to have the first byte be 0 though. Otherwise later on the whole machine won't be bootable anymore :). Any idea how to resolve this? The easy fix I could imagine would be to just not do the fixup on ARM, but that sounds slightly hackish :(. Alex -- To unsubscribe, e-mail: opensuse-arm+unsubscribe(a)opensuse.org To contact the owner, e-mail: opensuse-arm+owner(a)opensuse.org

2 3

[opensuse-arm] ARMv7 compile error
by Guillaume GARDET 02 Jul '12

02 Jul '12

|Hi,| || |||I made a package in OBS: https://build.opensuse.org/package/show?package=xf86-video-omap&project=hom… And I get this error for armv7: error: visibility attribute not supported in this configuration; ignored [-Werror=attributes] || cc1: all warnings being treated as errors| | I enabled armv5 for testing and it builds fine in armv5. I tried to play with some ./configure options with no effect. Any idea? Guillaume | -- To unsubscribe, e-mail: opensuse-arm+unsubscribe(a)opensuse.org To contact the owner, e-mail: opensuse-arm+owner(a)opensuse.org

3 7

[opensuse-arm] openSUSE openjdk 1.7 on ARM
by Alexander Graf 29 Jun '12

29 Jun '12

Hi Michal, Currently compiling openjdk on ARM fails for us with what looks like a generic zero engine issue. Do you have any idea what could be causing this? Did someone change internal interfaces and not update the zero vm? g++ -DLINUX -D_GNU_SOURCE -DCC_INTERP -DZERO -DTARGET_ARCH_NYI_6939861=1 -DARM -DZERO_LIBARCH=\"arm\" -DPRODUCT -I. -I/home/abuild/rpmbuild/BUILD/java-1_7_0-openjdk/openjdk/hotspot/src/share/vm/prims -I/home/abuild/rpmbuild/BUILD/java-1_7_0-openjdk/openjdk/hotspot/src/share/vm -I/home/abuild/rpmbuild/BUILD/java-1_7_0-openjdk/openjdk/hotspot/src/share/vm/precompiled -I/home/abuild/rpmbuild/BUILD/java-1_7_0-openjdk/openjdk/hotspot/src/cpu/zero/vm -I/home/abuild/rpmbuild/BUILD/java-1_7_0-openjdk/openjdk/hotspot/src/os_cpu/linux_zero/vm -I/home/abuild/rpmbuild/BUILD/java-1_7_0-openjdk/openjdk/hotspot/src/os/linux/vm -I/home/abuild/rpmbuild/BUILD/java-1_7_0-openjdk/openjdk/hotspot/src/os/posix/vm -I../generated -DHOTSPOT_RELEASE_VERSION="\"23.0-b21\"" -DHOTSPOT_BUILD_TARGET="\"product\"" -DHOTSPOT_BUILD_USER="\"abuild\"" -DHOTSPOT_LIB_ARCH=\"arm\" -DHOTSPOT_VM_DISTRO="\"OpenJDK\"" -DTARGET_OS_FAMILY_linux -DTARGET_ARCH_zero -DTARGET_ARCH_MODEL_zero -DTARGET_OS_ARCH_linux_zero -DTARGET_OS_ARCH_MODEL_linux_zero -DTARGET_COMPILER_gcc -fpic -fno-rtti -fno-exceptions -D_REENTRANT -fcheck-new -fvisibility=hidden -D_LITTLE_ENDIAN -pipe -g -O3 -fno-strict-aliasing -gstabs -DVM_LITTLE_ENDIAN -DINCLUDE_TRACE -Werror -Wpointer-arith -Wsign-compare -c -MMD -MP -MF ../generated/dependencies/cppInterpreter_zero.o.d -o cppInterpreter_zero.o /home/abuild/rpmbuild/BUILD/java-1_7_0-openjdk/openjdk/hotspot/src/cpu/zero/vm/cppInterpreter_zero.cpp /home/abuild/rpmbuild/BUILD/java-1_7_0-openjdk/openjdk/hotspot/src/cpu/zero/vm/cppInterpreter_zero.cpp: In static member function 'static void CppInterpreter::process_method_handle(oop, Thread*)': /home/abuild/rpmbuild/BUILD/java-1_7_0-openjdk/openjdk/hotspot/src/cpu/zero/vm/cppInterpreter_zero.cpp:817:7: error: 'get_ek_bound_mh_info' is not a member of 'MethodHandles' /home/abuild/rpmbuild/BUILD/java-1_7_0-openjdk/openjdk/hotspot/src/cpu/zero/vm/cppInterpreter_zero.cpp:965:7: error: 'get_ek_adapter_opt_swap_rot_info' is not a member of 'MethodHandles' The full log is available here once it's finished rebuilding again: https://build.opensuse.org/package/live_build_log?arch=armv7l&package=java-… Alex-- To unsubscribe, e-mail: opensuse-arm+unsubscribe(a)opensuse.org To contact the owner, e-mail: opensuse-arm+owner(a)opensuse.org

4 5

[opensuse-arm] introduction and bootstrap questions
by DuChene, StevenX A 28 Jun '12

28 Jun '12

Hello all: I have a couple of ARM systems here that have a four core chip by Marvell on a System On Module or SOM made by Cogent. This SOM is then socketed into a development motherboard that has four full size USB ports, two 1Gig-E network ports, two SATA ports, and a built-in serial-to-USB console. The SOM has 2GB of memory and the Marvell MV78460 chip has hardware floating point so the architecture shows up as armv7l or armv7hl or armv7hfp depending on what OS bits are loaded. I am supposed to be receiving a third system shortly for additional testing. Right now I have Fedora-13 ARM distro bits installed on a partition on a 500GB SATA hard drive. I was able to do this with a tarball of the root filesystem provided by the Fedora project. Once I had that unpacked on the hard drive I was able to alter the u-boot environment to still use the kernel resident on the internal NAND flash but use the root filesystem on the SATA disk. Later for ease of testing new kernel images I altered the u-boot environment again to use a kernel downloaded via tftp. I had downloaded the complete group of rpms Fedora-13 for ARM so I could setup a yum repo to facilitate adding more packages to the systems as needed. Unfortunately neither Marvell or Cogent are feeding any sort of platform support patches they have developed back to the standard kernel development process so I am limited to working with the source tree they have provided for right now. That is currently based around a 3.0.6 kernel tree. What I am wondering is if the OpenSUSE-arm project has a similar root filesystem tarball available that I could use to boot strap one of these systems to the current OpenSuSE-arm bits if they are released at all. That way I could do some testing of that build on this platform. -- Steven DuChene

2 1

[opensuse-arm] armv5 images
by Guillaume Gardet 28 Jun '12

28 Jun '12

Hi, would it be possible to enable JeOS image creation for armv5 so that we can have a rootfs for armv5 targets? I should receive my Raspberry Pi this week and I would like to test openSUSE armv5el on it. Cheers, Guillaume -- To unsubscribe, e-mail: opensuse-arm+unsubscribe(a)opensuse.org To contact the owner, e-mail: opensuse-arm+owner(a)opensuse.org

4 5

[opensuse-arm] xorg-x11-drv-omapfb config file
by Guillaume Gardet 27 Jun '12

27 Jun '12

Hi, Is there any reason to have all the lines commented in the config file 20-omapfb.conf for xorg-x11-drv-omapfb? Guillaume -- To unsubscribe, e-mail: opensuse-arm+unsubscribe(a)opensuse.org To contact the owner, e-mail: opensuse-arm+owner(a)opensuse.org

4 11